Why Statistics Can Be Misleading
Why Statistics Can Be Misleading
How Misleading Statistics Works | Whatagraph
Source: whatagraph.com
5 sources of misleading statistics (& how they can jeopardize your ...
Source: www.geckoboard.com
Common Types of Misleading Statistics in Advertising – And How to Spot Them
Source: www.singlegrain.com
is a common problem in data analysis and interpretation, affecting businesses, researchers, and policymakers worldwide.
Common Types of Misleading Statistics in Advertising – And How to Spot Them
Source: www.singlegrain.com
Sampling Biases
Top 5 Misleading Statistics in Software Development | HDWEBSOFT
Source: www.hdwebsoft.com
Statistics can be misleading when the sample used to represent a population is biased in some way. For example, if a survey is conducted among people who have internet access, it may not accurately represent the views of people who do not have internet access. This is because those who do not have internet access may have different opinions and demographics that are not captured in the survey. For instance, in a survey conducted by the Pew Research Center in 2020, the majority of respondents believed that social media had a positive impact on their lives. However, the survey only included people who had internet access and used social media regularly. This bias may have led to an inaccurate representation of the views of people who do not use social media or have limited internet access.
Misleading Statistics: How To Spot & Get Rid Of Them | Klipfolio
Source: www.klipfolio.com
Example of Sampling Bias
Misleading Statistics: How To Spot & Get Rid Of Them | Klipfolio
Source: www.klipfolio.com
| Social Media Users | Non-Social Media Users |
|---|---|
| 60% | 40% |
Misleading Statistics: How To Spot & Get Rid Of Them | Klipfolio
Source: www.klipfolio.com
This table shows a hypothetical example of a survey that finds that 60% of respondents use social media, while 40% do not. However, as mentioned earlier, the survey may be biased towards people who have internet access and use social media regularly. This means that the 40% of non-social media users may not accurately represent the views of people who do not use social media or have limited internet access.
Misleading Statistics: How To Spot & Get Rid Of Them | Klipfolio
Source: www.klipfolio.com
cherry-picking
Misleading Statistics: How To Spot & Get Rid Of Them | Klipfolio
Source: www.klipfolio.com
Statistics can also be misleading when researchers selectively present data that supports their hypothesis or claim. This is known as cherry-picking. For example, a researcher may present data that shows a correlation between two variables, but fail to mention that the correlation is weak or statistically insignificant. Cherry-picking can be done in various ways, including:
Misleading Statistics: How To Spot & Get Rid Of Them | Klipfolio
Source: www.klipfolio.com
- Selectively presenting data that supports a particular hypothesis or claim.
- Ignoring or downplaying data that contradicts a particular hypothesis or claim.
- Presenting data in a way that creates a misleading impression, such as using misleading or deceptive visuals.
statistics: how statistics can be misleading - Students | Britannica ...
Source: kids.britannica.com
Example of Cherry-picking
The Truth Behind the Numbers: Spotting Statistical Misuse | National ...
Source: blog.ons.gov.uk
Let's say a researcher wants to prove that a new diet program is effective in reducing weight. They may present data that shows that 90% of participants lost weight, but fail to mention that:
Misleading Beyond Visual Tricks: How People Actually Lie with Charts
Source: dl.acm.org
- Only 10% of participants actually completed the program.
- The weight loss was minimal and may not be clinically significant.
- The program was not sustainable in the long term.
Misleading Statistics Can Be Dangerous | PDF | Statistics | Survey ...
Source: www.scribd.com
Misleading Visualizations
Misleading Statistics - Andy Lutwyche
Source: andylutwyche.com
Statistics can also be misleading when visualizations are used in a way that creates a misleading impression. For example, a bar chart may be used to show a comparison between two groups, but the axis may be scaled in a way that makes one group appear larger than the other. Here are some common types of misleading visualizations:
Misleading Statistics and Customer Complexity – Donal Daly / 6 Rockets
Source: www.donaldaly.com
- 3D charts, which can make small changes appear large.
- Bar charts with a non-standard axis, which can create a misleading impression.
- Pie charts, which can make small percentages appear large.
Misleading Statistics Examples That Distort Reality
Source: examplesweb.net
Example of Misleading Visualization
How statistics can be misleading - Kidpid
Source: www.kidpid.com
| Country | Life Expectancy |
|---|---|
| Country A | 80 years |
| Country B | 70 years |
Misleading Statistics Examples: How Bad Statistics Misguides
Source: ninjatables.com
In this example, the life expectancy of Country A (80 years) appears to be much higher than that of Country B (70 years). However, if we were to use a log scale, the difference would be much smaller, and Country B's life expectancy would appear to be much closer to Country A's.
Misleading Statistics Examples: How Bad Statistics Misguides
Source: ninjatables.com
Raw Data
Misleading Statistics Examples: How Bad Statistics Misguides
Source: ninjatables.com
Statistics can be misleading when raw data is not presented in its entirety. For example, a study may present a summary statistic, such as the mean, without providing the underlying data. This can make it difficult to understand the spread of the data and may lead to misinterpretation. Here are some common reasons why raw data may be misleading:
Misleading Statistics Examples: How Bad Statistics Misguides
Source: ninjatables.com
- The data may be too large or complex to present in its entirety.
- The data may contain errors or inconsistencies.
- The data may be sensitive or confidential.
Misleading Statistics Examples: How Bad Statistics Misguides
Source: ninjatables.com
Example of Raw Data
Misleading Statistics Examples: How Bad Statistics Misguides
Source: ninjatables.com
| Variable | Value | | --- | --- | | Age | 25 | | Gender | Male | | Income | $50,000 | | Education | Bachelor's degree | | Job satisfaction | 7/10 | | Turnover intention | 3/10 | | Average commute time | 30 minutes | | Commute mode | Car | | Neighborhood satisfaction | 8/10 | In this example, the raw data presents a comprehensive picture of an individual's characteristics, behaviors, and attitudes. However, if we were to present only a summary statistic, such as the mean commute time, we may not get a full understanding of the underlying data.
Correlation Does Not Imply Causation
Statistics can be misleading when researchers confuse correlation with causation. For example, a study may find a correlation between two variables, but this does not necessarily mean that one variable causes the other. Here are some common reasons why correlation does not imply causation:
- There may be a third variable that causes both variables.
- There may be a reverse causality, where one variable causes the other.
- The correlation may be due to chance.
Example of Correlation Does Not Imply Causation
| Variable | Value | | --- | --- | | Smoking | 50% | | Heart disease | 20% | In this example, there may be a correlation between smoking and heart disease, but this does not necessarily mean that smoking causes heart disease. It is possible that there is a third variable, such as age or genetics, that causes both smoking and heart disease.